In this chapter, a brief history of Monte Carlo methods for finding the inverse of matrices is reviewed. For diagnoally dominant matrics, thress successful Monte Carlo methods: Plain Monte Carlo, Monte Carlo almost optimal and Sequential Monte Carlo are studied. Numberical algorithms based on such methods are given. To new (modified) algorithms: Tridiagonal Monte Carlo Almost Optimal and Sequential Monte Carlo without Augmented are presented. A comparison of the error and CPU time for these five Monte Carlo algorithms is done via numerical experiments in MATLAB. Finally, we give some concluding remarks according to the numerical results.